Eur. J. Immunol. 1990.20: 1911-1916
TSST-1 binding to murine cells
1911
Paul R. Schollo, Rafck-P. Sekalyv, Ana Diezo, Laurie H. Glimchern and Raif S. GehaO
Binding of toxic shock syndrome toxin-1 to murine major histocompatibility complex class I1 molecules*
Division of Allergy and Immunologyo, Children’s Hospital, and Department of Pediatricso, Harvard Medical School, Department of Cancer Biologyn, Harvard School of Public Health, Boston and Laboratoire d’Immunologiev, Institut de Recherches Cliiques de Mont&al, Montreal
The staphylococcal exotoxin toxic shock syndrome toxin-1 (TSST-1) has potent stimulatory effects on murine and human 1ymphocytes.This is the consequence of TSST-1 binding t o major histocompatibility complex (MHC) class I1 molecules and the engagement in a Vg-restricted fashion of the T cell receptor by the TSST-1-MHC class I1 complex. Using radioligand and functional assays we have recentlyshown that TSST-1binds t o all HLA-DR (n = 14), HLA-DQ (n = 2) and HLA-DP (n = 2) phenotypes tested. In this study,we have examined the ability of murine MHC class I1 molecules to bind TSST-1. Specific high-affinity binding of TSST-1 was detectable to unfractionated BALB-c (H-2d) and C57BL/6 (H-2b), but not to C3H (H-2k) spleen cells. The Kd of this binding estimated from Scatchard analysis was in the same nanomolar range as the Kd of binding of TSST-1 to HLA-DR. Binding of 1251-labeledTSST-1 to BALB/c-derived B cell lymphoma lines and to L cell transfectants correlated with the expression of I-A molecules, but not with the expression of I-E molecules. Furthermore, I-A+, I-E- cells but not I-A-, I-E+ cells were able to support TSST-1-induced Tcell proliferation. The binding affinity of TSST-1 for I-Ak appears to be much lower than for I-Ad. L cell transfectants expressing hybrid D R a : I-Ek molecules, but not those expressing I-Ei : DRlg molecules, could bind TSSY-1 and efficiently support TSST-1-induced Tcell proliferation. This suggests that minor differences in the highly homologous I-E, and DR, chains are critical in determining the affinity of the MHC class I1 molecule for TSST-1. These results demonstrate that the binding of TSST-1 to MHC class I1 molecules in the mouse, in contrast to humans, is strongly influenced by phenotype. Analysis of the molecular basis of these differences may help to localize staphylococcal exotoxin binding sites on MHC class I1 molecules.
1 Introduction Toxic shock syndrome toxin-1 (TSST-1) is a 22-kDa exotoxin produced by strains of Staphylococcus aureus which is implicated in the pathogenesis of toxic shock syndrome (TSS; [l-41). Like the structurally related staphylococcal enterotoxins (SE); [ 5 ] , TSST-1 has potent stimulatory effects on immune cells. These include the induction of proliferation and lymphokine secretion in rabbit, murine and human T lymphocytes [6], the activation of human B cells to Ig secretion [7], the induction of IL 1 and TNF by human monocytes [8-101 and the induction of hemopoietic growth factors in murine spleen cell cultures [ll]. Evidence has accumulated indicating that the mechanism of cellular activation by TSST-1 and related bacterial toxins involves a selective and specific interaction between the toxin and MHC class I1 molecules.Tcel1 activation by SE depends on the presence of MHC class 11+accessory cells, and mAb to MHC class I1 molecules block SE-induced Tcell activation [12-151. Specific high-affinity binding of SE to class I1 [I 82481
* This work was supported by National Institutes of Health grants
molecules has been demonstrated for TSST-1, SEA and SEB [14, 16-19]. Binding of TSST-1 and SEB was demonstrated to all 14 HLA-DR, 2 HLA-DQ and 2 HLA-DP phenotypes tested [20]. The Tcell response to SE, which is believed to occur upon the engagement of SE bound to class I1 molecules by the TcR, is restricted by TcR Vg phenotype [21,22] but not by MHC class I1 phenotype [17,20]. In addition, the formation of a ternary complex between SE, B cell MHC class I1 molecules and the TcR results in the delivery of a B cell activation signal which is MHC unrestricted [7]. Identification of the MHC class I1 determinants involved in SE binding is a prerequisite for understanding the mechanisms underlying these events. The finding that TSST-1 and SEB bind to human MHC class I1 molecules of diverse phenotypes suggests that these determinants are well conserved. In this study, we have examined the ability of murine MHC class I1 molecules to bind TSST-1.The results obtained show that the binding of TSST-1 to MHC class I1 molecules in the mouse, unlike in humans, is dependent on both isotype and allotype. Analysis of the molecular basis of these differences may help to localize SE binding sites on MHC class I1 molecules.
5T32H-D07321-03, AI20373-05 and 5P50-AI21163-07. Correspondence: Paul R. Scholl, M. D., Division of Allergy and Immunology, Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115, USA
2 Materials and methods
Abbreviations: SE: Staphylococcal exotoxins TSS: Toxic shock syndrome TSST-1: Toxic shock syndrome toxin-1
mAb MKD6 (anti-I-Ad), 14.4.4s (anti-I-Ed,k),11.4.1 (antiH-2k), L227 (anti-HLA-D) and OKMl (anti-CDllb) were
0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
2.1 mAb
0014-2980/90/0909-1911$3.50 + .25/0
1912
Eur. J. Irnmunol. 1990. 20: 1911-1916
F? R. Scholl, R.-F? Sekaly, A. Diez et al.
from hybridomas obtained from the American Type Culture Collection, Rockville, MD. mAb H39-487.7 (395, anti-I-Ak) and H7.8.26 (anti-I-Ek) were as described ~31.
2.2 Cell preparations
2.4 Prolieration assays Purified Tcells were cultured in 96-well plates, in a final volume of 200 p1 complete medium (human T cells) or RPMI 1640 medium with 2% FCS, 2 mhd L-glutamine, 100 U/ml penicillin and 50 mg/ml streptomycin (murine T cells), in the presence of accessory cells and toxins as indicated. Accessory cells were fixed by incubation in PBS containing 2% paraformaldehyde at room temperature for 15 min. Fixed cells were washed once in PBS, incubated in complete medium for 30 min at 37°C and washed twice more before being used in proliferation assays. In some experiments accessory cells were irradiated (10 OOO rad) instead of being fixed. Cultures were pulsed with [3H]dThd at 48 h and harvested at 72 h, and incorporation of [3H]dThd into DNA was determined as described [28].
The following murine L transfectants were used in this study: L57.23, expressing DR, :DRlg ([24]; the gift of Dr. R. Karr, University of Iowa), CA14.11.14 and CA36.1.3 expressing A; :A: and E i : E i ([25,26]; the gift of Dr. M. Pierres, Centre d'Immunologie, Marseille, France) and transfectants expressing the murinehuman mismatched MHC class I1 molecules DR,: I-Ei and I-Ei :DRlg (R. Lechler and R.-€? Sekaly, unpublished). The murine B cell lym homa lines M12.4.1 (I-Ad', I-Ed'), M12.A2 (I-Ad-, I-&, M12.B5 (I-Ad', I-Ed-) and M12.C3 (I-Ad-, I-Ed-) were as described [27]. Cell lines were grown in RPMI 1640 3 Results medium with 10% FCS, 2 mM L-glutamine, 100 U/ml penicillin and 50 mg/ml streptomycin (complete medium). 3.1 Binding of mI-labeled TSST-1 to murine spleen cells For L cell transfectants medium also contained G418 (250 mg/ml; Gibco, Grand Island, NY) or MXH (6 mg/ml Because of the ability of human MHC class I1 molecules to mycophenolic acid, 250 mg/ml xanthine and 15 mg/ml bind TSST-1, and because anti-MHC class I1 mAb have hypoxanthine) as appropriate for maintenance of selection. been shown to inhibit SEA- and SEB-induced proliferation Cultures were maintained at 37°C in humidified air con- of murine spleen cells, we investigated whether TSST-1 taining 5% C02. The surface expression of class II mole- might bind to murine as well as to human MHC class I1 cules was quantitated by indirect immunofluorescence as molecules. We first examined the ability of unfractionated previously described [HI. Cells were stained with the murine spleen cells to bind 1251-TSST-1.The results of appropriate anti-MHC class I1 mAb followed by FITC- binding experiments on three different mouse strains are conjugated goat anti-mouse IgG (Becton Dickinson, summarized in Table 1. 1251-TSST-1bound specifically to Mountain View, CA). Stained cells were fixed in 2% spleen cells of BALB/c (I-Ad', I-Ed') and C57BL/6 (I-Ab', paraformaldehyde and fluorescence determined on a FAC- I-Eb-) mice, but not detectably above background to spleen Scan analyzer (Becton Dickinson). cells of C3H (I-Ak', I-Ek') mice. Scatchard analysis of Mouse spleen cells were prepared by passing fresh spleens binding data was performed on cells from the two strains through steel mesh and washing the resulting cells in HBSS. that bound TSST-1, as illustrated in Fig. 1 for C57BL/6. Red cells were depleted by hypotonic lysis. Non-Tcells were Scatchard analysis gave values of approximately depleted by passage over nylon wool. The cells were then Kd = 20 nM. This is very close to the Kd for TSST-1 binding incubated with mAb MKD6 and 14.4.4S, washed and to human DR and DQ molecules [20]. incubated with goat anti-mouse IgG-coated magnetic beads (Advanced Magnetics Inc., Cambridge, MA). Antibody-coated cells were removed magnetically.The resultant preparations contained > 95% Tcells by FCM analysis and showed no proliferative response to TSST-1 in the absence lzoo of added accessory cells. h
Human T lymphocytes and monocytes were purified from PBMC as described previously [28]. Purification of T cells involved depletion of adherent cells by incubation on plastic petri dishes, rosetting with SRBC, passage over a nylon wool column and two cycles of antibody-complement lysis with mAb L227 and OKMl. The resultant preparations contained > 99% CD3+ cells, < 1% C D l l b+ cells and < 1% Bl f cells by FCManalysis.They routinely showed no proliferative response to PHA or Con A under conditions optimal for the proliferation of PBMC.
I000
1
E, v W
800
I 0
a c
600
I
c rn rn
400
ir
200
I-
o !
2.3 Toxin binding assays TSST-1 was provided by Dr. J. Parsonnet, Brigham and Women's Hospital, Boston, MA. TSST-1 used in some experiments was purchased from Toxin Technology (Madison, WI). Iodination of toxins and binding assays were as previously described [18-20].
k .
0
1
I
2.5~10-~
5x10-7
7 . 5 ~0-7 1
1xl 0-6
Cold TSST- 1 (M)
Figure I. Binding of TSST-1 to C57BL/6 spleen cells. Data points represent means (k SD) of triplicate determinations of 1251-TSST-1 (1 nM) bound to 1 x lo6unfractionated spleen cells in the presence of the indicated amounts of unlabeled TSST-1. Inset shows a Scatchard plot of the binding data.
Eur. J. Immunol. 1990. 20: 1911-1916
TSST-1 binding to murine cells
1913
Table 1. Binding of 1251-TSST-1to murine cells
mc class I1
Cells
phenotype Splenocytes BALBlc C57BLl6 C3H B cell lines M12.4.1 M12.B5 M12.A2 M12.C3
'251-TsST-1 bound (total)a)
I-Ad' I-Ed+ I-A~+',I-E-
3044f 62 2608f 59 I-A~+,I-E~+ 325f 18
I-A~+,I-E~+ I-A~+,I-EI-A-,I-E~+ I-A-,I-E-
9691 f 319 5663f 34 939f 59
Kd
'251-TSST-1 bound (nonspecific)
(nM)h)
276+ 46 2772: 20 2592: 36
29
15
a) Values represent means (+ SD) of triplicate determinations of the binding of 1251-TSST-1 M) to 1 X 1 8 unfractionated splenocytes or to 2.5 X 105 B cells. b) Values were derived by Scatchard analysis of single representative experiments. Similar results were obtained in two further experi-
-
1191f 56 738f 16 888 489 491+ 30
*
5 5 9 f 98
ments.
3.2 TSST-1 binds to I-Ad molecules, but not to I-Ed molecules In order to demonstrate unequivocally that the binding site for TSST-1 on murine cells is MHC class I1 molecules, we studied binding of 1251-TSST-1to M12.4.1, a lymphoma line derived from BALB/c mice that expresses I-Ad and I-E" molecules, and to three lines derived from M12.4.1 by mutagenesis and in which expression of I-Ad, I-Edor both has been deleted. As shown in Table 1, 1251-TSST-1bound to the parent line M12.4.1. In contrast, there was no detectable binding to the MHC class 11- mutant M12.C3. Furthermore, TSST-1 bound t o the I-Ad', I-Ed- line M12.B5 but not t o the I-Ad-, I-Ed' line M12.A2.The failure of binding to M12.A2 was not due to poor surface expression of I-Ed, as M12.A2 and M12.B5 expressed similar amounts of class I1 molecules as assessed by immunofluorescence (data not shown). These results indicate that murine MHC class I1 molecules, like their human counterparts, bind TSST-1, and suggest that in H-2d cells, binding is restricted to the I-A isotype. In the case of human class I1 molecules, we have previously shown that HLA-DP is able to support TSST-1-induced T
-
cell proliferation even though D P does not detectable bind TSST-1 in a radioligand assay [20]. This suggests that there is a low-affinity interaction between TSST-1 and DP, below the limit of detectability of the radioligand assay (Kd > 1 pM), but nevertheless sufficient to allow MHC class 11-dependent TSST-1-induced T cell proliferation. Therefore, in order to exclude the possibility of a similar low-affinity interaction betweenTSST-1 and I-Ed molecules we compared the capacity of I-Ad' and I-Ed' cells to support TSST-1-induced proliferation of human T cells. The ability of xenogeneic accessory cells t o support vigorous proliferation of human T cells induced by SE has been reported previously [29]. The results of a representative experiment are shown in Fig. 2. TSST-1 induced no proliferation of T cells in the presence of medium alone (not shown) or in the presence of the class 11- mutant, M12.C3. There was dose-dependent proliferation to TSST-1 in the presence of M12.4.1 (I-Ad', I-Ed'). The I-Ad', I-Ed- mutant M12.B5 supported TSST1-induced T cell proliferation as efficiently as M12.4.1.
-g
20000
20000
c
v
Y
10000
L
0
0
a
1 0
10000
L
n
0
c
20000 I
.-c
C
2 +
0, I
0 0 0
0.01
0.1
I
0
0.01
0.1
I
TSST-1 (ug/ml)
Figure 2. I-Ad molecules, but not I-Ed molecules, support TSST1-induced proliferation of human T cells. Values represent the mean (_+SD) of triplicate determinations of [3H]dThdincorporation by 1x 105 purified human Tcells cultured in the presence of accessory cells as indicated. Cultures were in medium alone (W) or with mAb MKD6 (63) or 14.4.4s ( ) added at a final concentration of 25 Fg/ml.
medium
M12.4.1
M12.C3
M12.A2
M12.B5
accessory cell Figure 3. LAd molecules, but not I-Ed molecules, support TSST1-induced proliferation of BALB/c mouse Tcells.Values represent the mean (_+SD) of triplicate determinations of [3H]dThd incorporation by 2 x 105 purified murineTcells cultured in the presence of 10 mg/ml TSST-1. Irradiated accessory cells were added at 5 x 103/well ( ) or 5 x 104/well (W). Values are corrected for background cpm in the absence of toxin.
1914
Eur. J. Immunol. 1990. 20: 1911-1916
P. R. Scholl, R.-I? Sekaly, A. Diez et al.
Fig. 2 also shows that the anti-I-Ad mAb MKD6 strongly inhibited TSST-1-induced T cell proliferation in the presence of both M12.B5 and M12.4.1. In contrast, the I-Ed-expressing cell M12.M failed to support TSST-1induced T cell proliferation at concentrations up to 100 ng/ml. At 1 pg/ml TSST-1, there was very weak proliferation with M12.A2; however, this may not be related to the expression of I-Ed on the accessory cell, as it was not blocked by the anti-I-Ed mAb 14.4.4s.
4000
It could be argued that the failure of M12.A2 to support TSST-1-induced Tcell proliferation might simply be due to a lack of a significant interaction between the mouse class I1 molecule I-Ed and the human TcR.We therefore performed a similar experiment using murine instead of human Tcells. As shown in Fig. 3,TSST-1 strongly induced the proliferation of murineTcells in the presence of either M12.4.1 or M12.BS cells. In contrast, only a very weak Tcell response toTSST-1 was seen in the presence of M12.C3 and M1 2 . M cells, and this was probably nonspecific, as it was not augmented by increasing the concentration of accessory cells. These results therefore confirm the absence of a functionally significant interaction between TSST-1 and I-Ed molecules.
1000
3.3 TSST-1 binds weakly to LAkmolecules, but not to I-ELmolecules
Having failed to detect binding of 1251-TSST-1to C3H (H-2k) splenocytes, we examined the ability of 1251-TSST-1 to bind to L cell transfectants expressing I-Ak and I-Ek molecules. The level of surface expression of MHC class I1 molecules by the I-Ak and I-Ek transfectants was similar as assessed by FCM (data not shown). Table2 shows that specific binding of 1251-TSST-1to the I-Ak transfectant, as demonstrated by the inhibition of binding by unlabeled TSST-1, was significantly (p < 0.01) but only weakly detectable above background. As well as being inhibited by cold toxin, binding was inhibited by the anti-I-A mAb 395, but not by mAb 14.4.4s. In contrast, binding of 1251-TSST-1to the I-Ek transfectant was nonspecific and was not significantly inhibited by cold TSST-1, nor by the anti-I-E mAb 14.4.4s. The I-Ak' and I-Ek' L cell transfectants were also tested for their ability to support TSST-1-induced Tcell proliferation.
3000
2000
0 0
10
1000
100
TSST-1 ng/ml
Figure 4 I-Ak molecules (O), but not I-Ek ( ) molecules, support TSST-1-induced proliferation of humanTcells.Values represent the mean (k SD) of triplicate determinations of [3H]dThd incorporation by 1 x 105purified humanTcells cultured in the presence of the indicated accessory cells. (W) DAP3.
As shown in Fig. 4, the I-Ak transfectant, but not the I-Ek transfectant or untransfected L cells, was able to support T cell proliferation to TSST-1, albeit to a lesser degree than the I-Ad' B cell lymphoma lines. Taken together, the above results clearly demonstrate preferential binding of TSST-1 to I-A molecules in both H-2d and H-2k strains. 3.4 TSST-1 binds to a hybrid DR,: I-Ep molecule
The failure of I-E molecules to bind TSST-1was a surprising finding, in view of the strong homology between I-E and HLA-DR molecules and in view of the ability of all allelic forms of D R tested (n = 14) [20] to bind TSST-1 with high affinity.The contribution of individual a and p chains of I-E and D R to TSST-1 binding was assessed by examining the binding of TSST-1 to L cells expressing hybrid class I1 molecules consisting of a murine a and human p chain, or vice versa. Both of these transfectants expressed similar quantities of the respective transfected MHC class I1 gene products as assessed by immunofluorescence staining and FCM (data not shown). Table 3 shows that s ecific binding of 1251-TSST-1was seen with the DR, : I-Ep transfectant,
P
Table 2. Binding of 1251-TSST-1to L cell transfectants expressing I-Ak or I-Ek moleculesa) Exp. no.
I
I1
III
cells DAP. 3 CA14.11.14 (2436.1.3 DAP.3 CA14.11.14 CA36.1.3 DAP.3 CA14.11.14 CA36.1.3
Medium
2092 k 106 3002f 17 1042f 33 879f 33 2066f 61 1042f 33 1217 f 110 1621f 43 811f 2
Cold TSST-1
1926 f 26 2193+53 1105 f 39 943 f 49 1632 f 64 1105 f 39 1262 It 42 m+10 796f 7
14.4.4s (anti-I-Ek)
ND 3054 f 207 1087f 45 ND 2095 f 93 1087 f45 ND ND ND
395 (anti-LAk)
ND
2082f67 1090f 50 ND 1499 f 72 1090 k 50 ND ND ND
a) Values represent the mean (fSD) of triplicate determinations of the binding of lZ5ITSST-1 (10nM) to 1 x 18, DAF! 3, CA14.11.14 (I-Ak') or CA36.1.3 (I-Ek+). Unlabeled TSST-1 was added at 1 x M. Purified mAb were added at 25 pg/ml.Values in bold represent significant inhibition of binding (p < 0.01, two-tailed t-test).
Eur. J. Immunol. 1990.20: 1911-1916
TSST-I binding to murine cells
1915
Table 3. Binding of 1251-TSST-1to L cells expressing species-mismatched MHC class IImolecules
Exp. no.
MHC class II phenotype
I
DR,, :DRlg I-EL::I-Ek DR,, :I-E,! I-Ek :DRlg
I1
m
W-TSW-1 bound laI-TSsT-l bound
(nonspecific)
(total)')
+
+
3313 129 2175 f 141 2002 f 374 1070 f 140
5968 NSb) 1844
DR,, :DRlp I-E: :I-EL DR, :I-E,f I-E: :DRlg
9281 225 22mi 46 3846 f 272 11m+ 10 5075 f 145 1673f 49 4136 k 134 653f 21
2290 f 188 1712f 76 16852 33 697f 14
2785
DR,, :DRlg I-EL::I-E' DR,, :I-E,f I-E: :DRlp
5092f 40 1976+ 68 4020 f 137 709f 74
1299f 1922f 1432f 65of
3793
61 71 53 64
but not with the I-Ei :DRlp transfectant. Likewise, Fig. 5 shows that the DR, :I-Ei transfectant efficiently supported TSST-1-induced proliferation of human T cells, whereas in the presence of the I-EL: :DRlp transfectant TSST-1 induced proliferation of T cells only weakly above background. These results suggest that the DR, chain is essential for binding and that DRp and I-Ep may contribute equally to TSST-1 binding.
4 Discussion The purpose of the present study was to examine the characteristics of binding of the staphylococcal exotoxin TSST-1 to murine MHC class11 molecules. The data demonstrate that TSST-1 binds with high affinity to two of three I-A molecules tested (I-Ab and I-Ad), and with weaker affinity to a third (I-Ak), but does not bind to two I-E molecules (I-Ed and I-Ek). h
E
n
u c
40000
0
'C
d
f
30000
0
n L 0 0
c
20000
IC
-0
i5
10000
2
I
*!
D (specific)
0 0
10
1 00
1000
TSST- 1 (ng/ml)
Figure5. I-Ep can pair with D& to form a hybrid molecule capable of supporting human T cell proliferation induced by TSST-1. (B) Medium; ( ) D%, :I-Ep; (H) I-E, :DRp. Results represent the mean (fSD) of triplicate determinations of [3H]dThdincorporation by 1X lo5 purified Tcells. Paraformaldehyde-fixed L cells were added at 1X I@/well.
NS
NS 245 1 NS
NS 2588 NS
a) Values represent means (& SD) of triplicate determinations of the binding of '251-TSST-1 to 5 x 105 L cell transfectants. b) Not significant.
Analysis of the role of the two murine Ia isotypes inTSST-1 binding was first performed on the B cell lymphoma line M12.4.1 and its mutagenized derivatives in which the expression of either I-Ad or I-Ed or both had been selectivelydeleted. 1251-TSST-1bound to the parent cell line and to the I-Ad', I-Ed- mutant, but not to the I-Ad-, I-Ed* mutant nor to the class I1 - mutant (Table 1). In addition, there was specific binding of lZI-TSST-1 to an L cell transfectant expressing I-Ak but not to a transfectant expressing I-Ek (Table 2). The differential ability of I-A vs. I-E molecules to bind TSST-1 was confirmed in a more sensitive functional assay which measures the capacity of cells bearing MHC class I1 molecules to support T cell proliferation toTSST-1. Both the I-Ad-expressingmurine B cell lymphoma lines as well as the I-Ak-transfected L cell supported TSST-1-induced Tcell proliferation (Figs. 2-4). In contrast, no significant proliferation was seen with I-Edor I-Ek-expressing cells. These results indicate that I-A molecules, but not I-E molecules, bind TSST-1. The ability to detect binding of 1251-TSST-1to I-Aktransfected L cells but not to spleen cells from C3H (H-2k) mice may be due to a combination of factors.These include the expression by splenocytes of relatively fewer MHC class I1 molecules and higher background binding to these cells.The similarity in intensity of I-A expression by spleen cells of the different strains tested as evaluated by immunofluorescence staining suggested that the poor binding of 1251-TSST-1 to I-Ak was due to lower affinity of this molecule for TSST-1, rather than to a lower number of binding sites. However,we did not attempt to measure the Kd directly. The difference in binding affinity for TSST-1 between I-Ad and I-Ak molecules contrasts with what has been reported in man, where there is little, if any, intraisotypic variability in the ability to bind TSST-1, SEA or SEB [17, 201. In contrast to the present results, a role for I-E molecules in the murine Tcell response to SEA and SEB was suggested by the demonstration that anti-I-A mAb strongly inhibited SEA- and SEB-induced activation of BALB/c spleen cells [15,30], and by the observation that I-E transfectants supported the proliferation of human T cells induced by
1916
F! R. Scholl, R.-P. Sekaly, A. Diez et al.
Eur. J. Immunol. 1990. 20: 1911-1916
SEA [29].This suggests that there are important differences 5 References in the ability of different staphylococcal exotoxins to bind 1 Davis, J. l?, Chesney, P. J.,Wand, l? J., Laventure, M. et al., N . to MHC class I1 molecules andor to stimulate a T cell Engl. J. Med. 1980. 303: 1429. response. Different exotoxins may bind differentially to 2 Bergdoll, M. S., Crass, B. A., Reiser, R. F., Robbins, R. N. and murine MHC class I1 molecules. This is suggested by a Davis, J. P., Lancet 1981. i: 1017. comparison of the present findings for TSST-1with those of 3 Schlievert, l? M., Shands, K. N., Dan, B. B., Schmid, G. F! and Buxser et al. [31] who detected binding of 1251-SEAand Nishimura, R.D., J. Infect. Dis. 1981. 143: 509. 1251-SEBto murine spleen cells, albeit only weakly above 4 Parsonnet, J., Gillis, Z. A., Richter, A. G. and Pier, G. F!, background and with a low estimated affinity ( K d in the Infect. lmmun. 1987. 55: 1070. micromolar range). In man, we have previously reported 5 Bergdoll, M. S . , in Jeljaszewicz, J. (Ed.), The Staphylococci, Gustav Fischer Verlag, New York 1985, p. 247. evidence that MHC class I1 molecules express two distinct 6 Poindexter, N. J. and Schlievert,l? M., J. Infect. D k . 1985.151: exotoxin-binding sites, which are differentially accessible to 65. TSST-1 and SEB, and that the affinity of human class I1 7 Mourad,W., Scholl, P., Dietz, A., Geha, R. S. and Chatila,T., J. molecules for SEB is an order of magnitude lower than that Med. 1989. 170: 2011. for TSST-1 [19]. Following their binding to MHC class I1 8 Exp. Ikejima,T., Dinarello, C. A., Gill, D. M. and Wolff, S. M., J. molecules, different toxins may activate distinct subsets of Clin. Invest. 1984. 73: 1312. Tcells in aVg-restricted fashion [22]. Different Vg TcR may 9 Parsonnet, J., Hickman, R. K., Eardley, D. P. and Pier, G. B., J. also have preferential affinity for ligands bound to different Infect. Dis. 1985. 151: 514. MHC class I1 molecules. If this is the case, then failure to 10 Parsonnet, J. and Gillis, Z. A., J. Infect. Dh. 1988. 158: respond toTSST-1 in the presence of a particular I-A or I-E 1026. phenotype may also reflect lack of the appropriately 11 Galelli, A., Anderson, S., Charlot, B. and Alouf, J. E., J. Immunol. 1989. 142: 2855. responsive Vg subset in the responding T cell population. However, it is unlikely that such a gap in theTcell repertoire 12 Fleischer, B. and Schrezenmeier, H., J. Exp. Med. 1988. 167: 1697. could explain our failure to detect a human Tcell response 13 Janeway, C. A. ,Yagi, J., Conrad, P. J., Katz, M. E., Jones, B., toTSST-1 in the presence of I-E molecules, because murine Vroegop, S. and Buxser, S., lmmunol. Rev. 1989. 107: 61. T cells also failed to respond t o TSST-1 in the context of 14 Fischer, H., Dohlsten, M., Lindvall, M., Sjogren, H.-0. and I-Ed. Our observation that I-E molecules fail to support Carlsson, R., J. lmmunol. 1989. 142: 3151. TSST-1-induced activation of murine T cells confirms 15 Vroegop, S. M. and Buxser, S. E., lnfect. Immun. 1989. 57: findings recently reported by another group [32]. 1816. The inability to detect binding of TSST-1 to I-E molecules in spite of the strong sequence homology between I-E and HLA-DR prompted us to examine the TSST-1 binding capacities of two hybrid DR/I-E molecules. Perhaps the most interesting finding of the present study was the observation that L cells transfected with DR, :I-E$, but not L cells transfected with I-Ei :DRlp, were able to bind 1251-TSST-1and to efficiently support TSST-1-induced Tcell proliferation. The marked contrast in TSST-1 binding between DR, :I-Ep and I-E, : I-Ep molecules suggests a major role for the MHC class I1 a chain inTSST-1 binding. This role could take one of several forms. One possibility is that TSST-1 binds exclusively to an a chain epitope expressed on DR,, but not on I-A,. An alternative possibility is that when co-expressed with either a DRg or I-Ep chain, D&, but not I-E, is able to impose a conformation on the fJ chain that is critical for the expression of aTSST-1 binding determinant. In either case, these findings suggest that minor differences in the highly homologous I-E, and DR, chains are critical in determining the affinity of the MHC class11 molecule €or TSST-1. Further studies, including the analysis of binding of TSST-1 to mutated MHC class I1 molecules, should clarify these issues, and are essential for a better understanding of the mechanisms by which the diverse cellular effects of SE are mediated. We thank M. Pierres for generously providing reagents.
Received January 8, 1990; in revised form May 4 , 1990.
16 Fraser, J. D., Nature 1989. 339: 221. 17 Mollick, J. A., Cook, R. G. and Rich, R. R., Science 1989.244: 817. 18 Scholl, l?, Diez, A., Mourad,W., Parsonnet, J., Geha, R. S. and Chatila, T., Proc. Natl. Acad. Sci. USA 1989. 86: 4210. 19 Scholl, P. R., Diez, A. and Geha, R. S., J. Immunol. 1989.143: 2583. 20 Scholl, I? R., Diez, A., Karr, R., Sekaly, R.-l?, Trowsdale, J. and Geha, R. S., J. lmmunol. 1990. 144: 226. 21 White, J., Herman, A., Pullen, A. M., Kubo, R., Kappler, J.W. and Marrack, P., Cell 1989. 56: 27. 22 Kappler, J., Kotzin, B., Herron, L., Gelfand, E.W., Bigler, R. D., Boylston, A., Carrel, S., Posnett, D. N., Choi, Y. and Marrack, P., Science 1989. 244: 811. 23 Pierres, M., Devaux, C., Dosseto, M. and Marchetto, S., lmrnunogenetics 1981. 14: 481. 24 Klohe, E. F!, Watts, R., Bahl, M., Alber, C., Yu, W.-Y, Anderson, R., Silver, J., Gregersen, P. K. and Karr, R. W., J. lmmunol. 1988. 141: 2158. 25 Malissen, B., Peele Price, M., Goverman, J., McMillan, M., White, J., Kappler, J., Marrack, l?, Pierres, A., Pierres, M. and Hood, L., Cell 1984. 36: 319. 26 Shastri, N., Malissen, B. and Hood, L., Proc. Natl. Acud. Sci. USA 1985. 82: 5885. 27 Griffith, I. J., Nabavi, N., Ghogawala, Z., Chase, C. G., Rodriguez, M., McKean, D. J. and Glimcher, L. H., J. Exp. Med. 1988. 167: 541. 28 Chatila, T. A., Schwartz, D. H., Miller, R. and Geha, R. S., Clin. lmmunol. Immunopathol. 1987. 44: 235. 29 Fleischer, B., Schrezenmeier, H. and Conradt, F!, Cell. Immunol. 1989. 120: 92. 30 Buxser, S. and Vroegop, S., J. lmmunogen. 1988. 15: 153. 31 Buxser, S., Bonventre, P. F. and Archer, D. L., Infect. Imrnun. 1981. 33: 827. 32 Uchiyama,T., Tadakuma,T., Imanishi, K., Araake, M., Saito, S. ,Yan, X.-J., Fujikawa, H., Igarashi, H. and Yamaura, N., J. lmmunol. 1989. 143: 3175.
TSST-1 binding to murine cells
1911
Paul R. Schollo, Rafck-P. Sekalyv, Ana Diezo, Laurie H. Glimchern and Raif S. GehaO
Binding of toxic shock syndrome toxin-1 to murine major histocompatibility complex class I1 molecules*
Division of Allergy and Immunologyo, Children’s Hospital, and Department of Pediatricso, Harvard Medical School, Department of Cancer Biologyn, Harvard School of Public Health, Boston and Laboratoire d’Immunologiev, Institut de Recherches Cliiques de Mont&al, Montreal
The staphylococcal exotoxin toxic shock syndrome toxin-1 (TSST-1) has potent stimulatory effects on murine and human 1ymphocytes.This is the consequence of TSST-1 binding t o major histocompatibility complex (MHC) class I1 molecules and the engagement in a Vg-restricted fashion of the T cell receptor by the TSST-1-MHC class I1 complex. Using radioligand and functional assays we have recentlyshown that TSST-1binds t o all HLA-DR (n = 14), HLA-DQ (n = 2) and HLA-DP (n = 2) phenotypes tested. In this study,we have examined the ability of murine MHC class I1 molecules to bind TSST-1. Specific high-affinity binding of TSST-1 was detectable to unfractionated BALB-c (H-2d) and C57BL/6 (H-2b), but not to C3H (H-2k) spleen cells. The Kd of this binding estimated from Scatchard analysis was in the same nanomolar range as the Kd of binding of TSST-1 to HLA-DR. Binding of 1251-labeledTSST-1 to BALB/c-derived B cell lymphoma lines and to L cell transfectants correlated with the expression of I-A molecules, but not with the expression of I-E molecules. Furthermore, I-A+, I-E- cells but not I-A-, I-E+ cells were able to support TSST-1-induced Tcell proliferation. The binding affinity of TSST-1 for I-Ak appears to be much lower than for I-Ad. L cell transfectants expressing hybrid D R a : I-Ek molecules, but not those expressing I-Ei : DRlg molecules, could bind TSSY-1 and efficiently support TSST-1-induced Tcell proliferation. This suggests that minor differences in the highly homologous I-E, and DR, chains are critical in determining the affinity of the MHC class I1 molecule for TSST-1. These results demonstrate that the binding of TSST-1 to MHC class I1 molecules in the mouse, in contrast to humans, is strongly influenced by phenotype. Analysis of the molecular basis of these differences may help to localize staphylococcal exotoxin binding sites on MHC class I1 molecules.
1 Introduction Toxic shock syndrome toxin-1 (TSST-1) is a 22-kDa exotoxin produced by strains of Staphylococcus aureus which is implicated in the pathogenesis of toxic shock syndrome (TSS; [l-41). Like the structurally related staphylococcal enterotoxins (SE); [ 5 ] , TSST-1 has potent stimulatory effects on immune cells. These include the induction of proliferation and lymphokine secretion in rabbit, murine and human T lymphocytes [6], the activation of human B cells to Ig secretion [7], the induction of IL 1 and TNF by human monocytes [8-101 and the induction of hemopoietic growth factors in murine spleen cell cultures [ll]. Evidence has accumulated indicating that the mechanism of cellular activation by TSST-1 and related bacterial toxins involves a selective and specific interaction between the toxin and MHC class I1 molecules.Tcel1 activation by SE depends on the presence of MHC class 11+accessory cells, and mAb to MHC class I1 molecules block SE-induced Tcell activation [12-151. Specific high-affinity binding of SE to class I1 [I 82481
* This work was supported by National Institutes of Health grants
molecules has been demonstrated for TSST-1, SEA and SEB [14, 16-19]. Binding of TSST-1 and SEB was demonstrated to all 14 HLA-DR, 2 HLA-DQ and 2 HLA-DP phenotypes tested [20]. The Tcell response to SE, which is believed to occur upon the engagement of SE bound to class I1 molecules by the TcR, is restricted by TcR Vg phenotype [21,22] but not by MHC class I1 phenotype [17,20]. In addition, the formation of a ternary complex between SE, B cell MHC class I1 molecules and the TcR results in the delivery of a B cell activation signal which is MHC unrestricted [7]. Identification of the MHC class I1 determinants involved in SE binding is a prerequisite for understanding the mechanisms underlying these events. The finding that TSST-1 and SEB bind to human MHC class I1 molecules of diverse phenotypes suggests that these determinants are well conserved. In this study, we have examined the ability of murine MHC class I1 molecules to bind TSST-1.The results obtained show that the binding of TSST-1 to MHC class I1 molecules in the mouse, unlike in humans, is dependent on both isotype and allotype. Analysis of the molecular basis of these differences may help to localize SE binding sites on MHC class I1 molecules.
5T32H-D07321-03, AI20373-05 and 5P50-AI21163-07. Correspondence: Paul R. Scholl, M. D., Division of Allergy and Immunology, Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115, USA
2 Materials and methods
Abbreviations: SE: Staphylococcal exotoxins TSS: Toxic shock syndrome TSST-1: Toxic shock syndrome toxin-1
mAb MKD6 (anti-I-Ad), 14.4.4s (anti-I-Ed,k),11.4.1 (antiH-2k), L227 (anti-HLA-D) and OKMl (anti-CDllb) were
0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
2.1 mAb
0014-2980/90/0909-1911$3.50 + .25/0
1912
Eur. J. Irnmunol. 1990. 20: 1911-1916
F? R. Scholl, R.-F? Sekaly, A. Diez et al.
from hybridomas obtained from the American Type Culture Collection, Rockville, MD. mAb H39-487.7 (395, anti-I-Ak) and H7.8.26 (anti-I-Ek) were as described ~31.
2.2 Cell preparations
2.4 Prolieration assays Purified Tcells were cultured in 96-well plates, in a final volume of 200 p1 complete medium (human T cells) or RPMI 1640 medium with 2% FCS, 2 mhd L-glutamine, 100 U/ml penicillin and 50 mg/ml streptomycin (murine T cells), in the presence of accessory cells and toxins as indicated. Accessory cells were fixed by incubation in PBS containing 2% paraformaldehyde at room temperature for 15 min. Fixed cells were washed once in PBS, incubated in complete medium for 30 min at 37°C and washed twice more before being used in proliferation assays. In some experiments accessory cells were irradiated (10 OOO rad) instead of being fixed. Cultures were pulsed with [3H]dThd at 48 h and harvested at 72 h, and incorporation of [3H]dThd into DNA was determined as described [28].
The following murine L transfectants were used in this study: L57.23, expressing DR, :DRlg ([24]; the gift of Dr. R. Karr, University of Iowa), CA14.11.14 and CA36.1.3 expressing A; :A: and E i : E i ([25,26]; the gift of Dr. M. Pierres, Centre d'Immunologie, Marseille, France) and transfectants expressing the murinehuman mismatched MHC class I1 molecules DR,: I-Ei and I-Ei :DRlg (R. Lechler and R.-€? Sekaly, unpublished). The murine B cell lym homa lines M12.4.1 (I-Ad', I-Ed'), M12.A2 (I-Ad-, I-&, M12.B5 (I-Ad', I-Ed-) and M12.C3 (I-Ad-, I-Ed-) were as described [27]. Cell lines were grown in RPMI 1640 3 Results medium with 10% FCS, 2 mM L-glutamine, 100 U/ml penicillin and 50 mg/ml streptomycin (complete medium). 3.1 Binding of mI-labeled TSST-1 to murine spleen cells For L cell transfectants medium also contained G418 (250 mg/ml; Gibco, Grand Island, NY) or MXH (6 mg/ml Because of the ability of human MHC class I1 molecules to mycophenolic acid, 250 mg/ml xanthine and 15 mg/ml bind TSST-1, and because anti-MHC class I1 mAb have hypoxanthine) as appropriate for maintenance of selection. been shown to inhibit SEA- and SEB-induced proliferation Cultures were maintained at 37°C in humidified air con- of murine spleen cells, we investigated whether TSST-1 taining 5% C02. The surface expression of class II mole- might bind to murine as well as to human MHC class I1 cules was quantitated by indirect immunofluorescence as molecules. We first examined the ability of unfractionated previously described [HI. Cells were stained with the murine spleen cells to bind 1251-TSST-1.The results of appropriate anti-MHC class I1 mAb followed by FITC- binding experiments on three different mouse strains are conjugated goat anti-mouse IgG (Becton Dickinson, summarized in Table 1. 1251-TSST-1bound specifically to Mountain View, CA). Stained cells were fixed in 2% spleen cells of BALB/c (I-Ad', I-Ed') and C57BL/6 (I-Ab', paraformaldehyde and fluorescence determined on a FAC- I-Eb-) mice, but not detectably above background to spleen Scan analyzer (Becton Dickinson). cells of C3H (I-Ak', I-Ek') mice. Scatchard analysis of Mouse spleen cells were prepared by passing fresh spleens binding data was performed on cells from the two strains through steel mesh and washing the resulting cells in HBSS. that bound TSST-1, as illustrated in Fig. 1 for C57BL/6. Red cells were depleted by hypotonic lysis. Non-Tcells were Scatchard analysis gave values of approximately depleted by passage over nylon wool. The cells were then Kd = 20 nM. This is very close to the Kd for TSST-1 binding incubated with mAb MKD6 and 14.4.4S, washed and to human DR and DQ molecules [20]. incubated with goat anti-mouse IgG-coated magnetic beads (Advanced Magnetics Inc., Cambridge, MA). Antibody-coated cells were removed magnetically.The resultant preparations contained > 95% Tcells by FCM analysis and showed no proliferative response to TSST-1 in the absence lzoo of added accessory cells. h
Human T lymphocytes and monocytes were purified from PBMC as described previously [28]. Purification of T cells involved depletion of adherent cells by incubation on plastic petri dishes, rosetting with SRBC, passage over a nylon wool column and two cycles of antibody-complement lysis with mAb L227 and OKMl. The resultant preparations contained > 99% CD3+ cells, < 1% C D l l b+ cells and < 1% Bl f cells by FCManalysis.They routinely showed no proliferative response to PHA or Con A under conditions optimal for the proliferation of PBMC.
I000
1
E, v W
800
I 0
a c
600
I
c rn rn
400
ir
200
I-
o !
2.3 Toxin binding assays TSST-1 was provided by Dr. J. Parsonnet, Brigham and Women's Hospital, Boston, MA. TSST-1 used in some experiments was purchased from Toxin Technology (Madison, WI). Iodination of toxins and binding assays were as previously described [18-20].
k .
0
1
I
2.5~10-~
5x10-7
7 . 5 ~0-7 1
1xl 0-6
Cold TSST- 1 (M)
Figure I. Binding of TSST-1 to C57BL/6 spleen cells. Data points represent means (k SD) of triplicate determinations of 1251-TSST-1 (1 nM) bound to 1 x lo6unfractionated spleen cells in the presence of the indicated amounts of unlabeled TSST-1. Inset shows a Scatchard plot of the binding data.
Eur. J. Immunol. 1990. 20: 1911-1916
TSST-1 binding to murine cells
1913
Table 1. Binding of 1251-TSST-1to murine cells
mc class I1
Cells
phenotype Splenocytes BALBlc C57BLl6 C3H B cell lines M12.4.1 M12.B5 M12.A2 M12.C3
'251-TsST-1 bound (total)a)
I-Ad' I-Ed+ I-A~+',I-E-
3044f 62 2608f 59 I-A~+,I-E~+ 325f 18
I-A~+,I-E~+ I-A~+,I-EI-A-,I-E~+ I-A-,I-E-
9691 f 319 5663f 34 939f 59
Kd
'251-TSST-1 bound (nonspecific)
(nM)h)
276+ 46 2772: 20 2592: 36
29
15
a) Values represent means (+ SD) of triplicate determinations of the binding of 1251-TSST-1 M) to 1 X 1 8 unfractionated splenocytes or to 2.5 X 105 B cells. b) Values were derived by Scatchard analysis of single representative experiments. Similar results were obtained in two further experi-
-
1191f 56 738f 16 888 489 491+ 30
*
5 5 9 f 98
ments.
3.2 TSST-1 binds to I-Ad molecules, but not to I-Ed molecules In order to demonstrate unequivocally that the binding site for TSST-1 on murine cells is MHC class I1 molecules, we studied binding of 1251-TSST-1to M12.4.1, a lymphoma line derived from BALB/c mice that expresses I-Ad and I-E" molecules, and to three lines derived from M12.4.1 by mutagenesis and in which expression of I-Ad, I-Edor both has been deleted. As shown in Table 1, 1251-TSST-1bound to the parent line M12.4.1. In contrast, there was no detectable binding to the MHC class 11- mutant M12.C3. Furthermore, TSST-1 bound t o the I-Ad', I-Ed- line M12.B5 but not t o the I-Ad-, I-Ed' line M12.A2.The failure of binding to M12.A2 was not due to poor surface expression of I-Ed, as M12.A2 and M12.B5 expressed similar amounts of class I1 molecules as assessed by immunofluorescence (data not shown). These results indicate that murine MHC class I1 molecules, like their human counterparts, bind TSST-1, and suggest that in H-2d cells, binding is restricted to the I-A isotype. In the case of human class I1 molecules, we have previously shown that HLA-DP is able to support TSST-1-induced T
-
cell proliferation even though D P does not detectable bind TSST-1 in a radioligand assay [20]. This suggests that there is a low-affinity interaction between TSST-1 and DP, below the limit of detectability of the radioligand assay (Kd > 1 pM), but nevertheless sufficient to allow MHC class 11-dependent TSST-1-induced T cell proliferation. Therefore, in order to exclude the possibility of a similar low-affinity interaction betweenTSST-1 and I-Ed molecules we compared the capacity of I-Ad' and I-Ed' cells to support TSST-1-induced proliferation of human T cells. The ability of xenogeneic accessory cells t o support vigorous proliferation of human T cells induced by SE has been reported previously [29]. The results of a representative experiment are shown in Fig. 2. TSST-1 induced no proliferation of T cells in the presence of medium alone (not shown) or in the presence of the class 11- mutant, M12.C3. There was dose-dependent proliferation to TSST-1 in the presence of M12.4.1 (I-Ad', I-Ed'). The I-Ad', I-Ed- mutant M12.B5 supported TSST1-induced T cell proliferation as efficiently as M12.4.1.
-g
20000
20000
c
v
Y
10000
L
0
0
a
1 0
10000
L
n
0
c
20000 I
.-c
C
2 +
0, I
0 0 0
0.01
0.1
I
0
0.01
0.1
I
TSST-1 (ug/ml)
Figure 2. I-Ad molecules, but not I-Ed molecules, support TSST1-induced proliferation of human T cells. Values represent the mean (_+SD) of triplicate determinations of [3H]dThdincorporation by 1x 105 purified human Tcells cultured in the presence of accessory cells as indicated. Cultures were in medium alone (W) or with mAb MKD6 (63) or 14.4.4s ( ) added at a final concentration of 25 Fg/ml.
medium
M12.4.1
M12.C3
M12.A2
M12.B5
accessory cell Figure 3. LAd molecules, but not I-Ed molecules, support TSST1-induced proliferation of BALB/c mouse Tcells.Values represent the mean (_+SD) of triplicate determinations of [3H]dThd incorporation by 2 x 105 purified murineTcells cultured in the presence of 10 mg/ml TSST-1. Irradiated accessory cells were added at 5 x 103/well ( ) or 5 x 104/well (W). Values are corrected for background cpm in the absence of toxin.
1914
Eur. J. Immunol. 1990. 20: 1911-1916
P. R. Scholl, R.-I? Sekaly, A. Diez et al.
Fig. 2 also shows that the anti-I-Ad mAb MKD6 strongly inhibited TSST-1-induced T cell proliferation in the presence of both M12.B5 and M12.4.1. In contrast, the I-Ed-expressing cell M12.M failed to support TSST-1induced T cell proliferation at concentrations up to 100 ng/ml. At 1 pg/ml TSST-1, there was very weak proliferation with M12.A2; however, this may not be related to the expression of I-Ed on the accessory cell, as it was not blocked by the anti-I-Ed mAb 14.4.4s.
4000
It could be argued that the failure of M12.A2 to support TSST-1-induced Tcell proliferation might simply be due to a lack of a significant interaction between the mouse class I1 molecule I-Ed and the human TcR.We therefore performed a similar experiment using murine instead of human Tcells. As shown in Fig. 3,TSST-1 strongly induced the proliferation of murineTcells in the presence of either M12.4.1 or M12.BS cells. In contrast, only a very weak Tcell response toTSST-1 was seen in the presence of M12.C3 and M1 2 . M cells, and this was probably nonspecific, as it was not augmented by increasing the concentration of accessory cells. These results therefore confirm the absence of a functionally significant interaction between TSST-1 and I-Ed molecules.
1000
3.3 TSST-1 binds weakly to LAkmolecules, but not to I-ELmolecules
Having failed to detect binding of 1251-TSST-1to C3H (H-2k) splenocytes, we examined the ability of 1251-TSST-1 to bind to L cell transfectants expressing I-Ak and I-Ek molecules. The level of surface expression of MHC class I1 molecules by the I-Ak and I-Ek transfectants was similar as assessed by FCM (data not shown). Table2 shows that specific binding of 1251-TSST-1to the I-Ak transfectant, as demonstrated by the inhibition of binding by unlabeled TSST-1, was significantly (p < 0.01) but only weakly detectable above background. As well as being inhibited by cold toxin, binding was inhibited by the anti-I-A mAb 395, but not by mAb 14.4.4s. In contrast, binding of 1251-TSST-1to the I-Ek transfectant was nonspecific and was not significantly inhibited by cold TSST-1, nor by the anti-I-E mAb 14.4.4s. The I-Ak' and I-Ek' L cell transfectants were also tested for their ability to support TSST-1-induced Tcell proliferation.
3000
2000
0 0
10
1000
100
TSST-1 ng/ml
Figure 4 I-Ak molecules (O), but not I-Ek ( ) molecules, support TSST-1-induced proliferation of humanTcells.Values represent the mean (k SD) of triplicate determinations of [3H]dThd incorporation by 1 x 105purified humanTcells cultured in the presence of the indicated accessory cells. (W) DAP3.
As shown in Fig. 4, the I-Ak transfectant, but not the I-Ek transfectant or untransfected L cells, was able to support T cell proliferation to TSST-1, albeit to a lesser degree than the I-Ad' B cell lymphoma lines. Taken together, the above results clearly demonstrate preferential binding of TSST-1 to I-A molecules in both H-2d and H-2k strains. 3.4 TSST-1 binds to a hybrid DR,: I-Ep molecule
The failure of I-E molecules to bind TSST-1was a surprising finding, in view of the strong homology between I-E and HLA-DR molecules and in view of the ability of all allelic forms of D R tested (n = 14) [20] to bind TSST-1 with high affinity.The contribution of individual a and p chains of I-E and D R to TSST-1 binding was assessed by examining the binding of TSST-1 to L cells expressing hybrid class I1 molecules consisting of a murine a and human p chain, or vice versa. Both of these transfectants expressed similar quantities of the respective transfected MHC class I1 gene products as assessed by immunofluorescence staining and FCM (data not shown). Table 3 shows that s ecific binding of 1251-TSST-1was seen with the DR, : I-Ep transfectant,
P
Table 2. Binding of 1251-TSST-1to L cell transfectants expressing I-Ak or I-Ek moleculesa) Exp. no.
I
I1
III
cells DAP. 3 CA14.11.14 (2436.1.3 DAP.3 CA14.11.14 CA36.1.3 DAP.3 CA14.11.14 CA36.1.3
Medium
2092 k 106 3002f 17 1042f 33 879f 33 2066f 61 1042f 33 1217 f 110 1621f 43 811f 2
Cold TSST-1
1926 f 26 2193+53 1105 f 39 943 f 49 1632 f 64 1105 f 39 1262 It 42 m+10 796f 7
14.4.4s (anti-I-Ek)
ND 3054 f 207 1087f 45 ND 2095 f 93 1087 f45 ND ND ND
395 (anti-LAk)
ND
2082f67 1090f 50 ND 1499 f 72 1090 k 50 ND ND ND
a) Values represent the mean (fSD) of triplicate determinations of the binding of lZ5ITSST-1 (10nM) to 1 x 18, DAF! 3, CA14.11.14 (I-Ak') or CA36.1.3 (I-Ek+). Unlabeled TSST-1 was added at 1 x M. Purified mAb were added at 25 pg/ml.Values in bold represent significant inhibition of binding (p < 0.01, two-tailed t-test).
Eur. J. Immunol. 1990.20: 1911-1916
TSST-I binding to murine cells
1915
Table 3. Binding of 1251-TSST-1to L cells expressing species-mismatched MHC class IImolecules
Exp. no.
MHC class II phenotype
I
DR,, :DRlg I-EL::I-Ek DR,, :I-E,! I-Ek :DRlg
I1
m
W-TSW-1 bound laI-TSsT-l bound
(nonspecific)
(total)')
+
+
3313 129 2175 f 141 2002 f 374 1070 f 140
5968 NSb) 1844
DR,, :DRlp I-E: :I-EL DR, :I-E,f I-E: :DRlg
9281 225 22mi 46 3846 f 272 11m+ 10 5075 f 145 1673f 49 4136 k 134 653f 21
2290 f 188 1712f 76 16852 33 697f 14
2785
DR,, :DRlg I-EL::I-E' DR,, :I-E,f I-E: :DRlp
5092f 40 1976+ 68 4020 f 137 709f 74
1299f 1922f 1432f 65of
3793
61 71 53 64
but not with the I-Ei :DRlp transfectant. Likewise, Fig. 5 shows that the DR, :I-Ei transfectant efficiently supported TSST-1-induced proliferation of human T cells, whereas in the presence of the I-EL: :DRlp transfectant TSST-1 induced proliferation of T cells only weakly above background. These results suggest that the DR, chain is essential for binding and that DRp and I-Ep may contribute equally to TSST-1 binding.
4 Discussion The purpose of the present study was to examine the characteristics of binding of the staphylococcal exotoxin TSST-1 to murine MHC class11 molecules. The data demonstrate that TSST-1 binds with high affinity to two of three I-A molecules tested (I-Ab and I-Ad), and with weaker affinity to a third (I-Ak), but does not bind to two I-E molecules (I-Ed and I-Ek). h
E
n
u c
40000
0
'C
d
f
30000
0
n L 0 0
c
20000
IC
-0
i5
10000
2
I
*!
D (specific)
0 0
10
1 00
1000
TSST- 1 (ng/ml)
Figure5. I-Ep can pair with D& to form a hybrid molecule capable of supporting human T cell proliferation induced by TSST-1. (B) Medium; ( ) D%, :I-Ep; (H) I-E, :DRp. Results represent the mean (fSD) of triplicate determinations of [3H]dThdincorporation by 1X lo5 purified Tcells. Paraformaldehyde-fixed L cells were added at 1X I@/well.
NS
NS 245 1 NS
NS 2588 NS
a) Values represent means (& SD) of triplicate determinations of the binding of '251-TSST-1 to 5 x 105 L cell transfectants. b) Not significant.
Analysis of the role of the two murine Ia isotypes inTSST-1 binding was first performed on the B cell lymphoma line M12.4.1 and its mutagenized derivatives in which the expression of either I-Ad or I-Ed or both had been selectivelydeleted. 1251-TSST-1bound to the parent cell line and to the I-Ad', I-Ed- mutant, but not to the I-Ad-, I-Ed* mutant nor to the class I1 - mutant (Table 1). In addition, there was specific binding of lZI-TSST-1 to an L cell transfectant expressing I-Ak but not to a transfectant expressing I-Ek (Table 2). The differential ability of I-A vs. I-E molecules to bind TSST-1 was confirmed in a more sensitive functional assay which measures the capacity of cells bearing MHC class I1 molecules to support T cell proliferation toTSST-1. Both the I-Ad-expressingmurine B cell lymphoma lines as well as the I-Ak-transfected L cell supported TSST-1-induced Tcell proliferation (Figs. 2-4). In contrast, no significant proliferation was seen with I-Edor I-Ek-expressing cells. These results indicate that I-A molecules, but not I-E molecules, bind TSST-1. The ability to detect binding of 1251-TSST-1to I-Aktransfected L cells but not to spleen cells from C3H (H-2k) mice may be due to a combination of factors.These include the expression by splenocytes of relatively fewer MHC class I1 molecules and higher background binding to these cells.The similarity in intensity of I-A expression by spleen cells of the different strains tested as evaluated by immunofluorescence staining suggested that the poor binding of 1251-TSST-1 to I-Ak was due to lower affinity of this molecule for TSST-1, rather than to a lower number of binding sites. However,we did not attempt to measure the Kd directly. The difference in binding affinity for TSST-1 between I-Ad and I-Ak molecules contrasts with what has been reported in man, where there is little, if any, intraisotypic variability in the ability to bind TSST-1, SEA or SEB [17, 201. In contrast to the present results, a role for I-E molecules in the murine Tcell response to SEA and SEB was suggested by the demonstration that anti-I-A mAb strongly inhibited SEA- and SEB-induced activation of BALB/c spleen cells [15,30], and by the observation that I-E transfectants supported the proliferation of human T cells induced by
1916
F! R. Scholl, R.-P. Sekaly, A. Diez et al.
Eur. J. Immunol. 1990. 20: 1911-1916
SEA [29].This suggests that there are important differences 5 References in the ability of different staphylococcal exotoxins to bind 1 Davis, J. l?, Chesney, P. J.,Wand, l? J., Laventure, M. et al., N . to MHC class I1 molecules andor to stimulate a T cell Engl. J. Med. 1980. 303: 1429. response. Different exotoxins may bind differentially to 2 Bergdoll, M. S., Crass, B. A., Reiser, R. F., Robbins, R. N. and murine MHC class I1 molecules. This is suggested by a Davis, J. P., Lancet 1981. i: 1017. comparison of the present findings for TSST-1with those of 3 Schlievert, l? M., Shands, K. N., Dan, B. B., Schmid, G. F! and Buxser et al. [31] who detected binding of 1251-SEAand Nishimura, R.D., J. Infect. Dis. 1981. 143: 509. 1251-SEBto murine spleen cells, albeit only weakly above 4 Parsonnet, J., Gillis, Z. A., Richter, A. G. and Pier, G. F!, background and with a low estimated affinity ( K d in the Infect. lmmun. 1987. 55: 1070. micromolar range). In man, we have previously reported 5 Bergdoll, M. S . , in Jeljaszewicz, J. (Ed.), The Staphylococci, Gustav Fischer Verlag, New York 1985, p. 247. evidence that MHC class I1 molecules express two distinct 6 Poindexter, N. J. and Schlievert,l? M., J. Infect. D k . 1985.151: exotoxin-binding sites, which are differentially accessible to 65. TSST-1 and SEB, and that the affinity of human class I1 7 Mourad,W., Scholl, P., Dietz, A., Geha, R. S. and Chatila,T., J. molecules for SEB is an order of magnitude lower than that Med. 1989. 170: 2011. for TSST-1 [19]. Following their binding to MHC class I1 8 Exp. Ikejima,T., Dinarello, C. A., Gill, D. M. and Wolff, S. M., J. molecules, different toxins may activate distinct subsets of Clin. Invest. 1984. 73: 1312. Tcells in aVg-restricted fashion [22]. Different Vg TcR may 9 Parsonnet, J., Hickman, R. K., Eardley, D. P. and Pier, G. B., J. also have preferential affinity for ligands bound to different Infect. Dis. 1985. 151: 514. MHC class I1 molecules. If this is the case, then failure to 10 Parsonnet, J. and Gillis, Z. A., J. Infect. Dh. 1988. 158: respond toTSST-1 in the presence of a particular I-A or I-E 1026. phenotype may also reflect lack of the appropriately 11 Galelli, A., Anderson, S., Charlot, B. and Alouf, J. E., J. Immunol. 1989. 142: 2855. responsive Vg subset in the responding T cell population. However, it is unlikely that such a gap in theTcell repertoire 12 Fleischer, B. and Schrezenmeier, H., J. Exp. Med. 1988. 167: 1697. could explain our failure to detect a human Tcell response 13 Janeway, C. A. ,Yagi, J., Conrad, P. J., Katz, M. E., Jones, B., toTSST-1 in the presence of I-E molecules, because murine Vroegop, S. and Buxser, S., lmmunol. Rev. 1989. 107: 61. T cells also failed to respond t o TSST-1 in the context of 14 Fischer, H., Dohlsten, M., Lindvall, M., Sjogren, H.-0. and I-Ed. Our observation that I-E molecules fail to support Carlsson, R., J. lmmunol. 1989. 142: 3151. TSST-1-induced activation of murine T cells confirms 15 Vroegop, S. M. and Buxser, S. E., lnfect. Immun. 1989. 57: findings recently reported by another group [32]. 1816. The inability to detect binding of TSST-1 to I-E molecules in spite of the strong sequence homology between I-E and HLA-DR prompted us to examine the TSST-1 binding capacities of two hybrid DR/I-E molecules. Perhaps the most interesting finding of the present study was the observation that L cells transfected with DR, :I-E$, but not L cells transfected with I-Ei :DRlp, were able to bind 1251-TSST-1and to efficiently support TSST-1-induced Tcell proliferation. The marked contrast in TSST-1 binding between DR, :I-Ep and I-E, : I-Ep molecules suggests a major role for the MHC class I1 a chain inTSST-1 binding. This role could take one of several forms. One possibility is that TSST-1 binds exclusively to an a chain epitope expressed on DR,, but not on I-A,. An alternative possibility is that when co-expressed with either a DRg or I-Ep chain, D&, but not I-E, is able to impose a conformation on the fJ chain that is critical for the expression of aTSST-1 binding determinant. In either case, these findings suggest that minor differences in the highly homologous I-E, and DR, chains are critical in determining the affinity of the MHC class11 molecule €or TSST-1. Further studies, including the analysis of binding of TSST-1 to mutated MHC class I1 molecules, should clarify these issues, and are essential for a better understanding of the mechanisms by which the diverse cellular effects of SE are mediated. We thank M. Pierres for generously providing reagents.
Received January 8, 1990; in revised form May 4 , 1990.
16 Fraser, J. D., Nature 1989. 339: 221. 17 Mollick, J. A., Cook, R. G. and Rich, R. R., Science 1989.244: 817. 18 Scholl, l?, Diez, A., Mourad,W., Parsonnet, J., Geha, R. S. and Chatila, T., Proc. Natl. Acad. Sci. USA 1989. 86: 4210. 19 Scholl, P. R., Diez, A. and Geha, R. S., J. Immunol. 1989.143: 2583. 20 Scholl, I? R., Diez, A., Karr, R., Sekaly, R.-l?, Trowsdale, J. and Geha, R. S., J. lmmunol. 1990. 144: 226. 21 White, J., Herman, A., Pullen, A. M., Kubo, R., Kappler, J.W. and Marrack, P., Cell 1989. 56: 27. 22 Kappler, J., Kotzin, B., Herron, L., Gelfand, E.W., Bigler, R. D., Boylston, A., Carrel, S., Posnett, D. N., Choi, Y. and Marrack, P., Science 1989. 244: 811. 23 Pierres, M., Devaux, C., Dosseto, M. and Marchetto, S., lmrnunogenetics 1981. 14: 481. 24 Klohe, E. F!, Watts, R., Bahl, M., Alber, C., Yu, W.-Y, Anderson, R., Silver, J., Gregersen, P. K. and Karr, R. W., J. lmmunol. 1988. 141: 2158. 25 Malissen, B., Peele Price, M., Goverman, J., McMillan, M., White, J., Kappler, J., Marrack, l?, Pierres, A., Pierres, M. and Hood, L., Cell 1984. 36: 319. 26 Shastri, N., Malissen, B. and Hood, L., Proc. Natl. Acud. Sci. USA 1985. 82: 5885. 27 Griffith, I. J., Nabavi, N., Ghogawala, Z., Chase, C. G., Rodriguez, M., McKean, D. J. and Glimcher, L. H., J. Exp. Med. 1988. 167: 541. 28 Chatila, T. A., Schwartz, D. H., Miller, R. and Geha, R. S., Clin. lmmunol. Immunopathol. 1987. 44: 235. 29 Fleischer, B., Schrezenmeier, H. and Conradt, F!, Cell. Immunol. 1989. 120: 92. 30 Buxser, S. and Vroegop, S., J. lmmunogen. 1988. 15: 153. 31 Buxser, S., Bonventre, P. F. and Archer, D. L., Infect. Imrnun. 1981. 33: 827. 32 Uchiyama,T., Tadakuma,T., Imanishi, K., Araake, M., Saito, S. ,Yan, X.-J., Fujikawa, H., Igarashi, H. and Yamaura, N., J. lmmunol. 1989. 143: 3175.
